Skip to content

GPU-Accelerated Binary Correspondence Search for Multishot Stereo Vision

License

LGPL-3.0, GPL-3.0 licenses found

Licenses found

LGPL-3.0
COPYING.LESSER
GPL-3.0
COPYING
Notifications You must be signed in to change notification settings

JMUWRobotics/libBICOS

Repository files navigation

libBICOS

BInary COrrespondence Search for multi-shot stereo imaging, with optional CUDA acceleration.

Citing:

This is the implementation of the corresponding paper to appear in Optical 3D Metrology. BibTeX will appear here when available.

Build:

Dependencies:

  • gcc or equivalent C compiler with C++20 support (build)
  • meson >= 1.1.0 (build)
  • opencv 4.x with cuda support
  • cuda 12.x including toolkit
  • cxxopts (optional, for cli)
  • benchmark (optional, for executing benchmarks)
$ git clone https://github.com/JMUWRobotics/libBICOS
$ cd libBICOS

# optional but recommended, pick a release depending on your needs
$ git checkout v1.2.0

$ meson setup builddir --buildtype release

# optional but recommended if you have access to a CUDA-capable GPU
$ meson configure -Dimplementation=cuda builddir

$ meson install -C builddir

The versioning scheme of the library is Semantic Versioning.

Usage

Linking

After installing, you can include /usr/local/include/BICOS/* and link against /usr/local/lib/libBICOS.so.

Commandline-interface

Alternatively, this project builds bicos-cli. To give you a feel for the parameters of BICOS, there is an example dataset provided in data with rectified imagery from a stereo camera, where images per side only differ in the projected light pattern. Calling:

$ bicos-cli data/{left,right} -q data/Q.cvstore --threshold 0.98 --stacksize 12 --variance 2.5 --step 0.1 -o /tmp/result.png

will get you:

Disparity

Example disparity

Pointcloud

Example pointcloud

While only requiring ~70ms (RTX4090) and ~1.5GB VRAM for matching on two stacks of 12 images each.

The most significant parameters can be summarized:

  • --threshold: increasing this reduces match-outliers. High values rely on a reliable pattern.
  • --variance: high values reduce coarse outliers, e.g. pixels where no pattern was projected. May reduce correct matches on low pattern contrast.
  • --step: optional value for subpixel-interpolation around matches.

Other settings are available; Try bicos-cli -h for details.

Benchmarking:

$ meson test --benchmark -C builddir --interactive

Light projection:

For starters, you may find https://github.com/Sir-Photch/VRRTest useful for projecting a light pattern using an aftermarket projector.

License

This library is licensed under the GNU Lesser General Public License v3.0 (or later). Please see COPYING and COPYING.LESSER for details.

LGPL-3.0-Logo